Gender and genre variation in weblogs

A relationship among language, gender, and discourse genre has previously been observed in informal, spoken interaction and formal, written texts. This study investigates the language/gender/genre relationship in weblogs, a popular new mode of computer-mediated communication (CMC). Taking as the dependent variables stylistic features identified in machine learning research and popularized in a Web interface called the Gender Genie, a multivariate analysis was conducted of entries from random weblogs in a sample balanced for author gender and weblog sub-genre (diary or filter). The results show that the diary entries contained more 'female' stylistic features, and the filter entries more 'male' stylistic features, independent of author gender. These findings problematize the characterization of the stylistic features as gendered, and suggest a need for more fine-grained genre analysis in CMC research. At the same time, it is observed that conventional associations of gender with certain spoken and written genres are reproduced in weblogs, along with their societal valuations.